دانلود کتاب AIML برای شبکه‌ها، سرویس‌ها و مدیریت: سیستم‌های خودمختار، مسیریابی، امنیت، AIOps، هوش مصنوعی عامل‌محور (Agentic AI)، LSO، MLFO و طراحی

Artificial Intelligence (AI) was born in 1950–1954 by Alan Turing’s publication of “Computer Machinery and Intelligence,” Arthur Samuel’s program to play checkers, and John McCarthy’s workshop at Dartmouth where he used the words “artificial Intelligence” first time. Artificial intelligence (AI) needs machine learning (ML) that uses algorithms like classification, clustering, and regression to parse data and learns from it. After that, ML can perform functions such as optimization and prediction. With the advancements in computing technologies in recent years, the development of AI/ML techniques have been accelerated tremendously. AI/ML has been impacting every industry from finance to healthcare. Public and private networks and services forming the communications infrastructure for all industries are also greatly impacted from the recent advances in computing and AI/ML technologies. AI-powered platforms that support evolving AI and networking demands and provide effective and scalable protection are claimed to be available by networking product vendors. AI is integrated into routers and networking solutions to create an AI-native ecosystem for autonomous networking, optimize performance, and simplify management. These are all good developments for the industry. Time will tell us whether these products are adequate to achieve full automation in networking or not and more importantly whether services supported by these products achieve full automation in coming years or not. Service automation is critical for the business. Days, weeks, or months for service ordering and implementation, failure identi_ication and repair, etc. are no longer acceptable. Operators need to reduce operational cost and react to changes in customer demands quickly while perfecting customer experience with their services. Automation of business services requires automated underlay (e.g., switches, routers, base stations, optical systems); automated overlay (e.g., Ethernet Virtual Connections—EVCs, IP Virtual Connections—IP VCs, SD-WAN Virtual Connections); automated management systems (e.g., OSS/BSS, orchestrator, subnetwork management system, element management system—EMS, Virtual Network Function Manager— VNFM, Network Functions Virtualization Infrastructure—NFVI, Virtualized Infrastructure Manager—VIM); and interoperability among AI/ML-powered underlay, overlay, and management systems requiring susbstantial amount of standardization. With AI/ML, we expect to achieve nearly self-managed networks/services. It may never be fully self-managed due to hardware and facilities connecting systems requiring human involvement for deployment and replacement. Connectivity and cloud providers have developed various procedures for automation over the years. These procedures can be starting point for AI/ML algorithms as described in the book. Generative AI models, such as Generative Adversarial Networks (GANs), Autoregressive models (ARs), and Variational Autoencoders (VAEs), can be used to generate synthetic data for training. This should accelerate modeling networks and services. In this book, I described my approaches as well as those in the literature for automating services, their underlays and management, after describing basics for AI/ML, service architectures, service attributes, and service OAM (operations, administration, maintenance). During the development of the book, it became very clear that tremendous amount of research and development are required for automating networking and services. I pointed some of them out in the book and encourage researchers in academia and industry to conduct research in these areas. The writing of this book took almost 2 years by working at nights and weekends. I want to thank my wife, Fusun, for her understanding and support. I would like to thank Nan Chen for his Foreword articulating well his vision and the content of the book. Also I would like to thank the Editor Mary James, the Editorial Assistant Kaitlyn Bonomo, and the Production Managers Thiyagarajan A. and Lokeshwaran Manickavasagam of Springer Publishing, who helped with the publication.